ITF roundtable reports quantifying the socio economic benefits of transport

Quantifying
the Socio-economic
Benefits of Transport

160

Roundtable Report


This work is published under the responsibility of the Secretary-General of the OECD. The
opinions expressed and arguments employed herein do not necessarily reflect the official
views of OECD member countries.
This document and any map included herein are without prejudice to the status of or
sovereignty over any territory, to the delimitation of international frontiers and boundaries
and to the name of any territory, city or area.

Please cite this publication as:
ITF (2017), Quantifying the Socio-economic Benefits of Transport, ITF Roundtable Reports, OECD
Publishing, Paris.
/>
ISBN 978-92-82-10808-6 (print)
ISBN 978-92-82-10809-3 (PDF)

Series: ITF Roundtable Reports
ISSN 2074-3378 (print)
ISSN 2074-336X (online)

The statistical data for Israel are supplied by and under the responsibility of the relevant Israeli authorities. The use
of such data by the OECD is without prejudice to the status of the Golan Heights, East Jerusalem and Israeli

settlements in the West Bank under the terms of international law.

Photo credits: Cover © Maciej Noskowski/iStockphoto.com.

Corrigenda to OECD publications may be found on line at: www.oecd.org/about/publishing/corrigenda.htm.

© OECD 2017
You can copy, download or print OECD content for your own use, and you can include excerpts from OECD publications, databases and
multimedia products in your own documents, presentations, blogs, websites and teaching materials, provided that suitable
acknowledgement of OECD as source and copyright owner is given. All requests for public or commercial use and translation rights
should be submitted to Requests for permission to photocopy portions of this material for public or commercial use shall
be addressed directly to the Copyright Clearance Center (CCC) at or the Centre français d’exploitation du droit de copie
(CFC) at


The International Transport Forum

The International Transport Forum is an intergovernmental organisation with 57 member countries. It
acts as a think tank for transport policy and organises the Annual Summit of transport ministers. ITF is
the only global body that covers all transport modes. The ITF is politically autonomous and
administratively integrated with the OECD.
The ITF works for transport policies that improve peoples’ lives. Our mission is to foster a deeper
understanding of the role of transport in economic growth, environmental sustainability and social
inclusion and to raise the public profile of transport policy.
The ITF organises global dialogue for better transport. We act as a platform for discussion and prenegotiation of policy issues across all transport modes. We analyse trends, share knowledge and promote
exchange among transport decision-makers and civil society. The ITF’s Annual Summit is the world’s
largest gathering of transport ministers and the leading global platform for dialogue on transport policy.
The Members of the ITF are: Albania, Armenia, Argentina, Australia, Austria, Azerbaijan, Belarus,
Belgium, Bosnia and Herzegovina, Bulgaria, Canada, Chile, China (People’s Republic of), Croatia,
Czech Republic, Denmark, Estonia, Finland, France, Former Yugoslav Republic of Macedonia, Georgia,

Germany, Greece, Hungary, Iceland, India, Ireland, Israel, Italy, Japan, Korea, Latvia, Liechtenstein,
Lithuania, Luxembourg, Malta, Mexico, Republic of Moldova, Montenegro, Morocco, the Netherlands,
New Zealand, Norway, Poland, Portugal, Romania, Russian Federation, Serbia, Slovak Republic,
Slovenia, Spain, Sweden, Switzerland, Turkey, Ukraine, the United Kingdom and the United States.
International Transport Forum
2, rue André Pascal
F-75775 Paris Cedex 16

www.itf-oecd.org

ITF Roundtable Reports
ITF Roundtable Reports present the proceedings of ITF roundtable meetings, dedicated to specific
topics notably on economic and regulatory aspects of transport policies in ITF member countries.
Roundtable Reports contain the reviewed versions of the discussion papers presented by international
experts at the meeting and a summary of discussions with the main findings of the roundtable. This work
is published under the responsibility of the Secretary-General of the OECD. The opinions expressed and
arguments employed herein do not necessarily reflect the official views of International Transport Forum
member countries. This document and any map included herein are without prejudice to the status of or
sovereignty over any territory, to the delimitation of international frontiers and boundaries and to the
name of any territory, city or area.



TABLE OF CONTENTS –

Table of contents

Executive summary................................................................................................................................ 9
Chapter 1. Improving transport cost-benefit analysis: Overview and findings ............................ 11
by Daniel Veryard

Introduction ................................................................................................................................. 12
Strategies to improve the practice and relevance of transport CBA ............................................ 13
Incorporating reliability benefits in CBA .................................................................................... 19
Incorporating wider economic impacts in CBA .......................................................................... 26
Notes ............................................................................................................................................ 33
References ................................................................................................................................... 35
Chapter 2. The valuation of travel-time variability ......................................................................... 39
by Mogens Fosgerau
Introduction ................................................................................................................................. 40
Some broader perspectives .......................................................................................................... 48
Estimating the parameters ........................................................................................................... 50
Conclusion ................................................................................................................................... 52
Notes ............................................................................................................................................ 54
References ................................................................................................................................... 55
Chapter 3. Forecasting travel-time reliability in road transport:
A new model for the Netherlands ............................................................................................. 57
by Marco Kouwenhoven and Pim Warffemius
Introduction ................................................................................................................................. 58
Methodology................................................................................................................................ 59
Data.............................................................................................................................................. 62
Testing alternative empirical relations for travel-time reliability ................................................ 65
A new empirical relation for the Netherlands.............................................................................. 70
Policy implications ...................................................................................................................... 74
Conclusions and future steps ....................................................................................................... 76
Acknowledgements ..................................................................................................................... 79
Notes ............................................................................................................................................ 79
References ................................................................................................................................... 80
Chapter 4. Estimating wider economic impacts in transport project prioritisation
using ex-post analysis ................................................................................................................ 83
by Glen Weisbrod

Introduction ................................................................................................................................. 84
Defining and measuring wider economic benefits and impacts .................................................. 85
Development of evidence-based planning and analysis methods in the US................................ 92
Use of case study and empirical analysis findings .................................................................... 101

QUANTIFYING THE SOCIO-ECONOMIC BENEFITS OF TRANSPORT — © OECD/ITF 2017

5


6 – TABLE OF CONTENTS

Implications for benefit accounting and decision support systems ........................................... 103
References ................................................................................................................................. 106
Chapter 5. Incorporating wider economic impacts within cost-benefit appraisal ...................... 109
by Anthony J. Venables
Introduction ............................................................................................................................... 110
The effects of a transport improvement ..................................................................................... 111
Proximity and productivity ........................................................................................................ 113
Investment and changes in land-use .......................................................................................... 115
Employment impacts ................................................................................................................. 118
Predicting quantity changes ....................................................................................................... 120
Summary and conclusions ......................................................................................................... 124
Notes .......................................................................................................................................... 125
References ................................................................................................................................. 126
Appendices.......................................................................................................................................... 128
Appendix I. Accessibility and productivity .............................................................................. 128
Appendix II. Investment and land-use change ......................................................................... 129
Participants list .................................................................................................................................. 131

Figures
Figure 1.1.
Figure 1.2.
Figure 1.3.
Figure 1.4.

Relationship between standard CBA and final economic effects of a transport project..... 12
Scope of CBA versus economic impact analysis................................................................ 16
Example of approach to estimating travel-time savings and reliability benefits ................ 20
Example travel time histogram (extreme events excluded three standard deviations
above the average) .............................................................................................................. 24
Figure 1.5. Framework for separating user benefits from wider economic impacts ............................. 26
Figure 2.1. Utility rates ......................................................................................................................... 42
Figure 2.2. Utility rates in the step model ............................................................................................. 44
Figure 2.3. Utility rates in the slope model ........................................................................................... 46
Figure 2.4. Theory and data .................................................................................................................. 51
Figure 3.1. The role of the transport models LMS/NRM and the post-processor LMS-BT in CBA .... 60
Figure 3.2. Travel-time distributions for four routes with different characteristics .............................. 64
Figure 3.3. Travel time per km versus standard deviation per km ........................................................ 65
Figure 3.4. Travel time versus standard deviation fitted with a linear function.................................... 66
Figure 3.5. Travel time per km versus standard deviation per km fitted with a linear relation and
a cubic polynomial .............................................................................................................. 67
Figure 3.6. Congestion index versus coefficient of variation fitted with a power law and with an
exponential function ........................................................................................................... 69
Figure 3.7. Travel time versus standard deviation fitted with a power law and a cubic polynomial .... 70
Figure 3.8. Mean delay versus standard deviation fitted with a combination of a linear and a
logarithmic function............................................................................................................ 71
Figure 3.9. Variability and delay relations for 250 routes for the morning peak .................................. 71
Figure 3.10. Results of fits for mean delay versus standard deviation under several choices
of the data analysis.............................................................................................................. 73

Figure 3.11. Mean delay versus standard deviation fitted with a linear function ................................... 73

QUANTIFYING THE SOCIO-ECONOMIC BENEFITS OF TRANSPORT — © OECD/ITF 2017


TABLE OF CONTENTS –

7

Figure 3.12. Ratio of travel-time benefits and reliability benefits for three projects, each for
two variants and two economic scenarios (high and low, figure adapted from 4Cast) ..... 76
Figure 4.1. Distinctions between CBA and EIA (US) .......................................................................... 85
Figure 4.2. Distinctions between CBA and EIA: Coverage of welfare and GDP effects (UK) ............ 86
Figure 4.3. Motivations for highway investments................................................................................. 94
Figure 4.4. Time lag in economic growth effects following highway investments .............................. 94
Figure 4.5. Relative concentration of industries by size of labour market ............................................ 97
Figure 4.6. Population concentration and manufacturing wage rates among counties
in central Appalachia .......................................................................................................... 98
Figure 4.7. Distribution of commuting time (cumulative per cent) .................................................... 100
Figure 5.1. The effects of a transport improvement ............................................................................ 112
Figure 5.2. Commercial development ................................................................................................. 117

Tables
Table 1.1.
Table 3.1.
Table 3.2.
Table 3.3.
Table 4.1.
Table 4.2.
Table 4.3.

Table 5.1.
Table A.1.

Approaches to valuing reliability benefits .......................................................................... 23
Characteristics of selected routes........................................................................................ 62
Best fit coefficients for the empirical relation between the standard deviation and the
mean delay (equation 9) for highway routes ...................................................................... 72
Best fit coefficients for the empirical relation between the standard deviation and the
mean delay (equation 9) for other routes ............................................................................ 74
Multi-criteria rating factors used for prioritisation (UK Appraisal Table is also
included for comparison) .................................................................................................... 90
Sensitivity of industries to access measures ....................................................................... 98
List of transport changes that are economic model inputs ................................................ 101
Predicting quantity changes .............................................................................................. 123
Accessibility and productivity .......................................................................................... 129

QUANTIFYING THE SOCIO-ECONOMIC BENEFITS OF TRANSPORT — © OECD/ITF 2017



EXECUTIVE SUMMARY –

9

Executive summary

Socio-economic cost-benefit analysis (CBA) provides a quantitative measure of the extent to which,
over its lifetime, a project or initiative will bring the community benefits that exceed its costs of
construction and operation. This report describes efforts to improve the quality of transport CBA and its
applicability to decision making. Three areas are addressed in detail: strategies for making the most of

CBA, valuing and forecasting reliability benefits, and capturing wider economic impacts. The report is
based on the papers and discussions at a roundtable meeting of 30 experts held in Paris in November
2015.
Key findings
CBA is a powerful framework that can be very useful to governments making investment decisions.
However the standard application of transport CBA faces three major challenges that have attracted the
attention of practitioners and researchers. First, there can be a mismatch between the information most
sought by decision makers (such as the impact on jobs and regional growth) and what is supplied by a
standard CBA (impact on national welfare and resource gains). Second, the scope of benefits captured in
a standard CBA is generally constrained by practical limitations of forecasting and valuation capability.
Third, fundamental changes to the quantity or locations of businesses, investments, households and
employment that are anticipated from transport investments are not captured within standard CBA.
Roundtable participants took the view that a multi-faceted approach is needed to address these shortfalls;
CBA theory and practice need to be gradually expanded to incorporate more impacts in the rigorous
valuation and forecasting framework; and CBA results need to be more effectively linked to other criteria
in the broader decision-making framework, including by bringing in a more diverse evidence base.
Main recommendations
CBA guidelines can be expanded to include reliability and some wider impacts
The current evidence base on the valuation and forecasting of reliability benefits, agglomeration
benefits and labour supply benefits provides a sufficiently rigorous basis for inclusion within the core
CBA of major transport projects. If properly applied, based on local evidence, the formal inclusion of
these benefits is better than either excluding them or applying simple mark-up rules.
Further research into reliability benefits is needed to improve confidence in results
There is significant variation among transport users in forming expectations about travel-time
reliability and in responding to it. Current approaches to valuing and forecasting reliability benefits take
a simplified approach in the interests of practicality. However, more research that disaggregates results
and examines the linkage between the standard of reliability and the transport choices made by users will
improve accuracy and build confidence in the results. The behavioural feedback can range from changes
in transport mode choices through to fundamental reorganisation of housing and business locations.
Closer international collaboration of researchers to share techniques, data and results is a promising

avenue for accelerating progress in this regard.

QUANTIFYING THE SOCIO-ECONOMIC BENEFITS OF TRANSPORT — © OECD/ITF 2017


10 – EXECUTIVE SUMMARY

Wider economic impacts should be examined in cases where they are expected to be significant
The relocation and reorganisation of businesses and households (a change in “economic geography”)
is a major motivation for some transport projects, such as regeneration schemes or transit-oriented
developments at rail stations. Wider economic impacts from such changes need to be communicated to
decision-makers. Formal inclusion of these impacts within CBA requires the identification of relevant
“market failures” in the project area (such as monopoly power) and the nature of the activity the project
affects. In many cases, there is no way of confidently predicting the economic geography outcomes when
co-ordinated actions among actors are required, such as in property development, so the best approach is
usually the application of scenarios. These scenarios can be informed by ex-post analysis of similar
projects as these can reveal success factors and the reasonable range of impacts that may materialise. For
any project though, a strong and scrutinised case for the inclusion of wider benefits is required to
overturn the traditional assumption that user benefits adequately account for the whole economic impact
of a project.
Further research into the impacts and tools for capturing wider impacts is needed
Practitioners currently face a choice between appraising the wider economic impacts of a transport
project by taking a “user benefits (CBA) plus wider benefits” approach or using a “big model” (such as a
Land Use-Transport Interaction model) to capture all impacts. Roundtable participants generally
favoured the former approach. This is because no big models are yet able to adequately capture all the
relevant impacts of a transport project, they require a very large amount of data and the complexity of the
modelling required undermines transparency. However, given that LUTI or general equilibrium model, at
least in theory, should be able to produce answers to the most relevant questions these may ultimately be
best placed to address the limitations of CBA. With further research, there may come a point where the
big models become responsive, accurate, and cheap enough to apply as the preferred project appraisal

approach. Most roundtable participants though were of the view that that time has not yet arrived, so
there is a strong motivation to keep improving the practice of CBA.
CBA can play an important role in decision making, but need not dominate
CBA is valuable yet imperfect. Appraisal is most useful to decision-makers when the CBA approach
is clearly aligned with the objectives sought, when it draws on the best local evidence available and when
the shortfalls and uncertainties are clearly highlighted in the analysis. Available evidence will not always
be of sufficient quality to justify inclusion within the formal CBA. In such cases, supporting frameworks
and alternative evidence will be useful to communicate possible project impacts. Two options in
particular were highlighted at the roundtable. First, by drawing quantitative and qualitative insights from
similar past projects, ex-post analysis can give vivid insights into potential economic geography changes
and their driving forces. Second, complementary tools, such as economic impact analysis and qualitative
explanation of non-quantifiable impacts, can help address shortfalls inherent in CBA. Presenting such
diverse information to decision-makers is better than producing a single performance measure, since the
latter can generally only be achieved either by including bold and unfounded assumptions or by ignoring
impacts altogether.

QUANTIFYING THE SOCIO-ECONOMIC BENEFITS OF TRANSPORT — © OECD/ITF 2017


1. IMPROVING TRANSPORT COST-BENFIT ANALYSIS: OVERVIEW AND FINDINGS –

11

Chapter 1

Improving transport cost-benefit analysis:
Overview and findings
Daniel Veryard1

Socio-economic cost-benefit analysis (CBA) is a powerful framework that can be very useful to governments

making investment decisions. However the standard application of transport CBA has room for improvement. This
chapter describes efforts to improve the quality of transport CBA and its applicability to decision making. Three
areas are addressed in detail: strategies for making the most of CBA, valuing and forecasting reliability benefits,
and capturing wider economic impacts. The chapter is based on the papers and discussions at a roundtable
meeting of 30 experts held in Paris in November 2015. Roundtable participants took the view that a multi-faceted
approach is needed to address the shortfalls; CBA theory and practice need to be gradually expanded to
incorporate more impacts in the rigorous valuation and forecasting framework; and CBA results need to be more
effectively linked to other criteria in the broader decision-making framework, including by bringing in a more
diverse evidence base.

1

International Transport Forum, Paris, France

QUANTIFYING THE SOCIO-ECONOMIC BENEFITS OF TRANSPORT — © OECD/ITF 2017


12 – 1. IMPROVING TRANSPORT COST-BENFIT ANALYSIS: OVERVIEW AND FINDINGS

Introduction

For most transport economists working in the field of public policy making the preferred tool for
project prioritisation and selection is socio-economic cost-benefit analysis (CBA). The standard practice
of CBA involves the estimation and valuation of direct benefits to transport users when the transport
system is incrementally improved. Some direct external impacts are also often accounted for, such as
impacts on congestion and the environment. Under the assumptions of constant returns to scale and
perfect competition, this framework (if well applied) captures the ultimate economic impacts of
expanded production, wages and employment (Dodgson, 1973; Jara-Diaz, 1986). That is, under these
assumptions, the shaded area labelled “standard CBA scope” in Figure 1.1 is fully consistent with the
lower box in the figure (“transmitted economic effects”).

Figure 1.1. Relationship between standard CBA and final economic effects of a transport project

However, the standard application of transport CBA faces challenges that have attracted the attention
of practitioners and researchers. These can be described with further reference to Figure 1.1 and broadly
fall into three related themes:

QUANTIFYING THE SOCIO-ECONOMIC BENEFITS OF TRANSPORT — © OECD/ITF 2017


1. IMPROVING TRANSPORT COST-BENFIT ANALYSIS: OVERVIEW AND FINDINGS –

13

•

Relevance – There is often a mismatch between the information wanted by decision makers
compared to what is supplied by a standard CBA. For instance, CBA supplies measures of
resource benefits and social welfare benefits from the perspective of the nation. But decision
makers may wish to understand the final (transmitted) impacts on jobs and economic activity in
their region (the lower box in Figure 1.1). CBA results are not able to be directly interpreted in
terms of final economic impacts, except in the broadest terms.

•

Sophistication – The scope of benefits captured in a standard CBA is generally constrained by
practical limitations of forecasting and valuation capability, even within the grey shaded
“standard CBA scope” area in Figure 1.1. In this chapter we discuss in particular the limitations
and recent improvements in the estimation of reliability benefits.

•

Coverage – Transport improvements can encourage fundamental changes to the quantity or
locations of businesses, investments, households and employment that are not captured within
standard CBA (the right hand side of Figure 1.1). When, as is often the case, the theoretical
assumptions underpinning standard CBA don’t fully apply, the equivalence between direct
benefits and final economic impacts breaks down. There are genuinely additional benefits that
conceptually need to be added to a CBA to be complete. These wider economic impacts are
discussed further in this chapter.

In relatively straightforward transport investment decisions, such as choosing between alignment
options for roads, standard CBA is likely to provide sufficient relevant evidence. However, there will be
cases where relying only on the estimation of impacts that are “easily” forecast and valued will not be
sufficient. Focusing only on well-established benefits, such as direct travel-time savings to existing users,
could bias investment towards projects that are well suited to narrow assessments (e.g. road expansion)
and disadvantage projects that address particular objectives such as measures to encourage people to shift
to public transport or strategies to address freight travel-time variability and could miss the critical
aspects of projects designed to have a more “transformative” effect such as regeneration of depressed
regions or enhancing the potential for growth through, for example, deepening and thickening the labour
market.1
Progress has been made in the past decade to improve the quality of CBA and its applicability to
decision making. This report reviews the state of the art in two areas in particular: reliability benefits and
wider economic impacts. It summarises four papers on these issues commissioned for a roundtable
meeting, held in Paris in November 2015, and brings together the discussions among leading experts at
the roundtable. Edited versions of the four input papers comprise the remaining chapters of this report.

Strategies to improve the practice and relevance of transport CBA

Transport CBA is a powerful framework that provides a quantitative measure of the extent to which,
over its lifetime, a project or initiative will bring the community benefits that exceed its costs of
construction and operation. The framework is sufficiently flexible to be used to support a wide range of

decisions. For instance, CBA can be used to filter out poor projects from consideration, or can be used to
optimise a relatively promising project (e.g. refining alignments). The particular role of CBA can depend
on the quality of the portfolio of transport projects that come under consideration (ITF, 2011).

QUANTIFYING THE SOCIO-ECONOMIC BENEFITS OF TRANSPORT — © OECD/ITF 2017


14 – 1. IMPROVING TRANSPORT COST-BENFIT ANALYSIS: OVERVIEW AND FINDINGS

The role that CBA plays in the overall decision making about transport investment is also affected by
the relevance of the criticisms that CBA does not capture all of a project’s expected impacts, or is unable
to provide all the information that is relevant to decision makers. In some jurisdictions, such as several
US states, CBA is not mandated or considered at all in decisions over transport investments (Weisbrod,
Chapter 4). In most countries, and particularly in northern Europe and Australasia, CBA is a central,
although not always dominant, part of the overall decision-making framework (Mackie and Worsley,
2013).
CBA results tend to be supplemented with other quantitative and qualitative information when
presented to a decision maker. Where CBA is applied, the quantified (and monetised) estimates of
project benefits are typically presented in a business case alongside descriptions of impacts that are more
difficult to value (such as heritage), and information about how the direct (user) impacts of a project are
transmitted through the economy into changes in employment and output. The latter effects can be
estimated using a range of economic modelling techniques discussed later in this section. In jurisdictions
where CBA is not applied, the local and regional economic impact estimates (as opposed to national
welfare benefits) are given more prominence in decision making. Regardless of the technical approach to
quantifying expected impacts, it is almost always the case that the final decision is taken based on a
judgement over quantitative and qualitative information (Mackie and Worsley, 2013).
Roundtable participants identified and discussed a range of alternative and complementary strategies
that could be pursued to improve the quality of transport CBA and its usefulness for decision makers:
•

improved strategic alignment and communication of results

•

applying complementary appraisal frameworks to capture effects outside traditional CBA

•

drawing evidence from previous projects (case studies and ex-post analysis)

•

extending the toolkit and scope of accepted CBA practice

•

tailoring each CBA to the project’s context and objectives.

Improved strategic alignment and communication
Ideally projects should be proposed on the basis of a careful strategic planning exercise that starts
from the overall objectives or mission for a jurisdiction (supranational, national, regional or local).
Projects are, however, often the result of more “instinctive” proposals from politicians or public officials.
Several participants at the roundtable reported having been asked to appraise a project for which there
was no clear statement of what problem it was trying to address, or what objectives it was trying to
achieve. In such circumstances, where there is no “narrative”, it is difficult to predict and advise on
whether the project is likely to be a success, regardless of the analytical framework applied.
Practitioners can apply the standard CBA framework to estimate travel-time savings, safety benefits
and environmental improvements regardless of the nature of the project and its objectives. However, it is
unlikely that the results on their own will be sufficiently meaningful for political decision makers or their
constituents. Instead, there is a need to first place the project into the overall strategic context and to

develop a qualitative case for which transport, social, environmental and economic variables the project
is likely to have the greatest effects.
Clarifying the strategic intentions of the project allows the CBA to align the assessment of benefits
with the achievement of objectives. Clarity and alignment with project objectives will not only ensure the
CBA is relevant to the project but will also ensure the results can be explained to decision makers in
relevant language, with conclusions that are relevant to the project’s original objectives.

QUANTIFYING THE SOCIO-ECONOMIC BENEFITS OF TRANSPORT — © OECD/ITF 2017


1. IMPROVING TRANSPORT COST-BENFIT ANALYSIS: OVERVIEW AND FINDINGS –

15

Applying complementary appraisal frameworks
An obvious strategy to overcome the limitations of CBA is to complement it with an alternative
appraisal framework. Two broad categories were discussed at the roundtable: economic impact analysis,
which measures changes in business revenue, profits, wages and jobs for a specified area; and scoring
methods, such as multi-criteria analysis (MCA).
Economic impact analysis can be used for ex-ante assessment to directly forecast the final economic
impacts in the lower box of Figure 1.1 above. Economic impact analysis uses macroeconomic methods,
such as econometric regressions, computable general equilibrium (CGE) models, spatial CGE models or
mesoscopic models, and aims to forecast the increase in output or employment from an increase in the
stock or quality of transport infrastructure (see Chapter 4 or Vickerman (2008) for a review). Economic
impact analysis takes a different perspective to CBA but is somewhat overlapping in scope (Figure 1.2).
Participants at the roundtable recognised the potential value of these models for two primary tasks:
communicating economic impacts in terms relevant to decision makers (jobs and economic activity), and
for highlighting the regional and socio-economic distribution of impacts.
CGE models in particular have the capability to account for some of the market imperfections that
fall outside the scope of traditional CBA and can be used to provide information on the extent to which

resources are displaced from one location to another (discussed further in the final section of this chapter)
and more generally on who benefits and who loses from the investment. However, CGE models are
data-intensive, costly to run and are difficult to critique due to their mathematical complexity. The
question is whether the transport assets in such macro models could be specified with sufficient accuracy
to account for improvements in dimensions such as travel-time variability and end-to-end public
transport journeys that are relevant in many contemporary transport policy decisions.
Weisbrod (Chapter 4) cites two systems regional of economic models used by some US state
transport departments for ex-ante assessment of projects. These CGE-type formulations do incorporate
many detailed transport characteristics as inputs, and provide outputs of regional economic impacts. As
an alternative to this system of models approach, one participant argued that the theoretical foundations
of stand-alone CGE models could be re-built to include a more detailed representation of the transport
system and its quality in consumer and producer functions. While this approach would be useful for
understanding the effects of major transport-sector technological improvements, other participants
argued that most transport projects under consideration are more marginal, so marginal appraisal
approaches are sufficiently accurate.2 Economic impact analysis methods were therefore considered most
appropriate as a (non-additive) complement to CBA for major projects (or programmes of projects)
where non-marginal effects on the economy are expected and where the expense of the modelling is
justifiable.

QUANTIFYING THE SOCIO-ECONOMIC BENEFITS OF TRANSPORT — © OECD/ITF 2017


16 – 1. IMPROVING TRANSPORT COST-BENFIT ANALYSIS: OVERVIEW AND FINDINGS

Figure 1.2. Scope of CBA versus economic impact analysis

Source: Adapted from Weisbrod (Chapter 4).

Scoring methods, such as multi-criteria analysis (MCA) or scorecards, require an analyst to assign
scores that assess the expected achievement of the project against pre-defined objectives. In the various

US states, transport-related objectives can be inferred from transport departments’ mission statements,
with strong emphasis given to economic development, environmental improvement, mobility and safety
(Volpe Transportation Systems Center, 2012). In MCA, objectives are generally weighted according to
their priority for the decision maker (and to an extent their degree of overlap with other objectives).
While guidelines are available to support MCA decision processes, economists are generally sceptical of
its application due to issues of double-counting, arbitrary implicit valuations,3 lack of viability threshold,
and susceptibility to “gaming” (Dobes and Bennett, 2009).
Weisbrod (Chapter 4) argues that the level of government making the transport investment decision
can influence the choice of appraisal framework. For example, national decision making in the UK has
traditionally preferred CBA over CGE and MCA approaches as it focuses more on the overall efficiency
of investments as a use of national funds, with less emphasis on the distributional impacts that
approximately “net out” across winners and losers.4 In contrast, where decisions are made at state or
regional level, these local decision makers are very concerned with the potential for redistribution of
economic activity into their jurisdiction, but are less concerned if these gains are made at the expense
of other jurisdictions. CGE and other economic impact analysis models may therefore support this kind
of decision making (though some participants noted that CBA can also be used to answer at least some of
the questions about the spatial distribution of benefits). As some member countries devolve decision
making from national to lower levels of government, such considerations may come further into focus
(Mackie and Worsley, 2013).
The question of trust and empowerment was also raised in the context of the choice of appraisal
framework. The choice between MCA and CBA can be viewed as a question of who the community
prefers to make judgements on their behalf. A community that has strong trust that elected politicians and
their planners will make trade-offs in the public interest might prefer the simplicity, comprehensiveness
and immediacy of MCA. In contrast, a community that is suspicious that an MCA might be manipulated
for political objectives that do not align with social welfare might prefer the reassurance of the more
technocratic approach of CBA.
MCA can be useful for bringing together expected impacts that cannot be appropriately valued in
CBA, such as irreversible heritage and environmental effects, alongside those effects that can be valued
with CBA techniques (Weisbrod, Chapter 4). Several participants argued that in such instances, it is
preferable to instead highlight these relevant impacts qualitatively alongside the CBA result, as is the

QUANTIFYING THE SOCIO-ECONOMIC BENEFITS OF TRANSPORT — © OECD/ITF 2017


1. IMPROVING TRANSPORT COST-BENFIT ANALYSIS: OVERVIEW AND FINDINGS –

17

practice in the UK’s Appraisal Summary Table, rather than adding a false sense of precision with the
integrated result of the MCA approach, particularly given concerns about double counting.
Drawing evidence from previous projects
CBAs of major urban transport projects usually estimate the project’s future impacts using strategic
transport forecasting models. Even with major advances in the modelling techniques applied in strategic
models since they were developed in the 1950s, there are still many real-world behaviours of firms and
households that are not reflected in these models and hence the CBAs that rely solely on them. A crucial
missing element is the redistribution and reorganisation of businesses and households that might happen
in response to a significant improvement in accessibility in a with-project case (the right-hand side of
Figure 1.1). In other words, in standard CBA transport activity develops under the “fixed land use”
assumption that ignores changes in “economic geography”. But changes in economic geography are a
major motivation for some transport projects, such as regeneration schemes or transit-oriented
developments at new rail stations. In such instances, evidence on these changes is critical for:
•

describing to decision makers the extent to which resources and activity may be
redistributed across the economy, and what factors might influence these outcomes

•

understanding the mechanisms and magnitudes of the economic geography changes that can
give rise to very specific additional benefits not already captured in standard CBA

(discussed in the final section of this chapter).

Roundtable participants discussed two strands of research that provide evidence from previous
projects to explain the potential impacts of a proposed project: case studies and ex-post statistical
analysis. Both types of approaches seek to infer relationships between transport improvements (the top of
Figure 1.1) and their final impacts on the regional or national economy (the bottom of Figure 1.1).
Case studies
Weisbrod (Chapter 4) discusses the US approach to learning from previously implemented projects.
Evidence is sought on the many economic and social effects that occur after different types of projects
have been implemented, such as the development of industry clusters in areas with improved
accessibility. Over 100 case studies are available in the EconWorks database hosted by the AASHTO, the
association of state transportation departments. Across the cases, different forms of accessibility are
found to be important to each industry sector. For example, professional services require a large
commuter catchment and an international airport, while manufacturing businesses need to be able to
access markets and suppliers in a one-day truck turn-around. Analysis of the wide range of projects and
contexts also allows different types of clustering to be identified with respect to its supporting transport
infrastructure.
One cluster was discussed at the roundtable in some detail: a linear automobile supply chain cluster
along the I-65 and I-75 highways in Kentucky and Tennessee. Here, suppliers can traverse the full extent
of the cluster by truck in a single day to allow “just-in-time” supply chains and “lean production”
processes (Weisbrod, Chapter 4). The rural location simultaneously gives access to urban markets and to
low cost local labour. One roundtable participant stressed that although this may be an interesting result
of the transport infrastructure, it was not necessarily delivering a social welfare gain above alternative
spatial patterns. It may even be possible that by allowing employers to exercise monopsony power in
local labour markets there could be a social loss. This kind of granular insight is important to the
narrative presented to decision makers. It illustrates the risks inherent in scaling up the results of
economic impact analysis from the firm or specific area scale to drawing conclusions on socio-economic
welfare changes from a national perspective.

QUANTIFYING THE SOCIO-ECONOMIC BENEFITS OF TRANSPORT — © OECD/ITF 2017

18 – 1. IMPROVING TRANSPORT COST-BENFIT ANALYSIS: OVERVIEW AND FINDINGS

Ex-post statistical analysis
There are significant difficulties in using case study evidence in ex-ante analysis of other projects.
Chief among the challenges is the difficulty of separating the effect of a specific project from other
factors that would help explain the development outcomes following a transport investment. Empirical
ex-post analysis of US case study data attempts to control for these factors by asking the analyst to
identify a counter-factual reference case (e.g. state-wide average changes over the study period) and to
qualitatively attribute (less than 100% of) the observed changes in the study area (relative to the
reference case) to the project (Weisbrod, Chapter 4). However, the data gathered and techniques
recommended are generally not yet sufficient to do so in a statistically robust way.
A recent roundtable on ex-post analysis identified several challenges that need to be overcome to be
able to robustly attribute observed outcomes to a particular transport project or initiative (ITF,
Forthcoming). The first challenge is that, unlike controlled trials in health or education, the location that
receives a “treatment” (a transport investment) is not randomly selected from a set of options. Instead,
investments tend to be made precisely where for example congestion or crash rates are the worst, or
economic activity is supressed. A second challenge in ex-post transport analysis is that each transport
project takes place within a different transport network and socio-economic context. This means that an
appropriate counterfactual or “control” cannot be identified or specified to allow for the effects of the
project to be neatly identified (Worsley, 2014).
Several, but not all, participants at the present roundtable were optimistic that the data and tools
required to delineate the effects of projects from other factors could be brought together in the near term.
Ex-post data being collected – for example, in the US (Weisbrod, Chapter 4) and France (Bonnafous,
2014) – continues to expand in quantity and quality (though some participants were sceptical that the
projects selected for analysis may not always be a neutral selection of successes and failures). Graham
(2014) describes the use of statistical inference methods to remove the influence of “confounding
effects” by simulating a random assignment of investments across alternative locations. These techniques
have been applied with success in the ex-post analysis of road safety projects. “Crash modification

functions” have been developed for different project types and contexts that allow safety impacts from a
project to be estimated ex-ante (ITF, 2012). It may therefore be possible in the future to get to the same
point for economic impacts from transport projects: the development of a range of “economic impact
factors” for different project types and contexts that could be applied in ex-ante assessments.
Extending the toolkit and scope of accepted CBA practice
In northern Europe at least, the majority of effort by researchers to improve the usefulness of CBA
has been to extend the framework, rather than to replace it with an alternative (e.g. UK, France and
Sweden). This can be considered as either adding to, or improving on, the items in “direct resource
benefits”, “welfare benefits” or “other resource benefits” in Figure 1.1.
Several participants emphasised that the “burden of proof” required to incorporate additional effects
in CBA is very high, with significant scepticism from national oversight bodies, such as Treasury
departments. In practice, the inclusion of a “new” benefit to the accepted CBA framework requires
researchers and practitioners to robustly demonstrate that the additional effect or benefit:
(1) is theoretically additional to other benefits captured in CBA
(2) can be valued in a way that is robust and does not overlap with the valuation of related effects
that are already included
(3) can be adequately forecast, with and without the project intervention.

QUANTIFYING THE SOCIO-ECONOMIC BENEFITS OF TRANSPORT — © OECD/ITF 2017


1. IMPROVING TRANSPORT COST-BENFIT ANALYSIS: OVERVIEW AND FINDINGS –

19

The most topical of the extensions to the CBA framework in recent years are those classified as
“wider economic benefits” or “wider impacts”. Before treating these in detail (in the final section), we
will discuss reliability benefits (the next section), whose additionality is less theoretically contested
(point 1 above), but are challenging to forecast and value (points 2 and 3).
Tailoring each CBA to the project’s context and objectives

The discussion at the roundtable emphasised the diversity of possible impacts from major transport
projects (and how these could vary by location) as well as the complexity of the tools that can be
deployed to forecast and value these appropriately in CBA. A natural question arises: is it better to have a
standard toolkit (or model) that is applied equally to every project assessed by a government agency, or
should some parts of the toolkit be deployed only where the relevant benefits were expected to be
significant for the project? The majority view on this question was that a scalable and modular approach
was best in practice, though there were reservations to this view. This issue will be taken up further in the
final section.

Incorporating reliability benefits in CBA

Travel-time variability is inherent in all modes of transport. This variability often leads to significant
costs to travellers and the economy. One of the recurring themes of the roundtable discussions was the
diversity of the experiences of travellers with unreliability: from the mild inconvenience of arriving
unexpectedly late at a holiday destination, through to complete spoilage of time-critical freight or missed
business meetings. Measures that improve reliability are therefore certain to be valuable to the
community, so the ability to effectively include reliability benefits into transport CBA is paramount to
ensure that projects that particularly improve reliability are properly reflected in project prioritisation.
The diversity of trip purposes and responses to variability make its measurement and valuation
particularly difficult. Roundtable discussions centred on the three items required for reliability to be
included in CBA, as noted in the previous section. That is, we need first a clear definition and measure of
reliability that does not overlap with other items in CBA; then a unit value for the cost of unreliability
(i.e. the benefit of an improvement in reliability); and finally an approach to forecasting reliability with
and without a project intervention. These three points can be demonstrated with reference to an example
(Figure 1.3). In the figure, additionality of average travel-time benefits and reliability benefits is ensured
if the valuation parameters (on the left hand side) are distinctly estimated in the same study. The
approach to forecasting future variability described later in the section relies on a relation between the
average levels of delay or congestion and travel-time variability. In the example, reliability benefits are
80% of the magnitude of average travel-time savings.
The consensus view at the roundtable was that reliability is a critical dimension of transport system

performance that should be included within CBA if a sufficient evidence base was available to value and
forecast improvements. Implemented approaches range from relatively sophisticated valuation and
forecasting approaches (e.g. the UK and France), to percentage mark-ups on the average travel-time
savings (e.g. currently in the Netherlands) to excluding it altogether. Outside the US, where the focus has
been on freight reliability, much of the research focus has been on road transport for passenger travel.
Extensions of the research to cover public transport were also discussed at the roundtable.

QUANTIFYING THE SOCIO-ECONOMIC BENEFITS OF TRANSPORT — © OECD/ITF 2017


20 – 1. IMPROVING TRANSPORT COST-BENFIT ANALYSIS: OVERVIEW AND FINDINGS

Figure 1.3. Example of approach to estimating travel-time savings and reliability benefits

Valuation parameters
Value of average
travel time savings
(e.g. $10/hour)

Benefits

Forecasts of impacts

(

Average travel time
without project
(e.g. 40 minutes)

Average travel time

with project
(e.g. 34 minutes)

)

Travel time savings
(10 x (40-34)/60 = $1)

)

Reliability savings
(8 x (10-4)/60 = $0.80)

Statistical relation between
delays and variability
Value of travel time
reliability
(e.g. $8/hour of σ)

(

Travel time variability
without project
(e.g. σ = 10 minutes)

Travel time variability
with project
(e.g. σ = 4 minutes)

Defining reliability

The definition of travel-time reliability is not straightforward compared with the definition of average
travel time. While the latter is widely understood by the public and is a concept routinely estimated by
strategic transport models, the same cannot be said for reliability. Three elements of the definition of
reliability were discussed at the roundtable: the nature of unreliability, travellers’ perceptions of this
unreliability, and responses to unreliability.
The nature and causes of unreliability
Data on travel times in many different contexts demonstrates systematic variability across the day,
week and year (ITF, 2010). These variations are generally due to widely understood peaks and troughs in
passenger demand relating to work and education schedules. On a given day, other variations in travel
times can be due to somewhat regular occurrences, such as traffic signals, rain, crashes or network
maintenance. A third type of cause is what can be called “extreme events”, such as flooding, severe
incidents or network closures (de Jong and Bliemer, 2015). Roundtable participants were somewhat
divided on whether a hard distinction should be drawn between more routine types of travel-time
variability and that caused by extreme events. Put another way, the question is whether to separately treat
outliers (very long travel times) of travel-time distributions when we estimate our reliability metric (or
metrics).
How travellers perceive unreliability
In distinguishing between average travel-time savings and travel-time reliability savings, we are
necessarily allowing travellers some imperfection in their ex-ante knowledge of travel-time outcomes. At
one extreme, with perfect knowledge of their travel times for upcoming trips, transport projects could
produce only average travel-time savings since there would be no ex-ante uncertainty about travel time.
In practice, travellers will have some knowledge of the likely travel time for their upcoming trip, but they
will recognise there is some risk of the actual travel time being longer than this “likely” time. That is,
travellers implicitly have a distribution of travel times in mind for their trip before the trip is completed.
The discussion at the roundtable raised a number of questions about when these expectations are formed
and what information they are based upon. Two main sources of information about upcoming travel
times are:
•

personal experience travelling on the same route or service

QUANTIFYING THE SOCIO-ECONOMIC BENEFITS OF TRANSPORT — © OECD/ITF 2017


1. IMPROVING TRANSPORT COST-BENFIT ANALYSIS: OVERVIEW AND FINDINGS –

•

21

external information about current travel conditions, such as traffic reports, “smart”
navigation on smartphones or GPS units, real-time information signage en route, or even
weather reports.

The difficulty in defining an ex-ante measure of reliability is that different people will be armed with
widely differing information about the likely travel time associated with essentially the same trip. For
example, two people may be about to travel on the same stretch of highway departing at 8:20 am on
Tuesday. The first is from out of town and has only a paper map to help direct him to his destination. The
second person regularly commutes along the highway, has reviewed the traffic incident reports over
breakfast and she found yesterday that the week-long strike on the rail network meant that travel times
were higher than usual for a Monday. The challenge is to relate the expectations from such diverse
travellers to observable information about travel-time variability, as well as available information about
how travellers value improvements in reliability.
Responses to unreliability
A further aspect that is relevant in defining reliability is the ways that travellers may respond to
unreliability. One response to expected unreliability of an upcoming trip is for a traveller to build in an
additional time margin above her expected (point estimate) travel time and depart slightly early
(Fosgerau, Chapter 2). People may choose to change their departure time altogether to a period where
travel times are more predictable. Travellers with repeated experience in using a route or service that find
it to be unacceptably unreliable may change route, service or travel mode. Even once the trip is

underway, if a traveller finds the trip is taking longer than was expected, they can sometimes change
route – especially if they have real-time information (either from road-side signs or an on-board device).
People and businesses may make more fundamental changes in response to unreliability (a change in
“economic geography” discussed in the final section below). For example, if a regular commuter finds
travel times too unreliable, she may move house to be closer to her workplace or closer to a more reliable
transport node, such as a rail station. Enterprises have an even wider range of choices in response to
unreliable transport services. While initially, late-running deliveries can result in service penalties and
driver overtime costs, freight companies may decide to purchase additional fleet (and pay additional
drivers) to meet customer service levels at a given level of network reliability. More fundamentally still,
businesses may choose to reorganise their entire supply chains and production processes to be more local
(Weisbrod, Chapter 4), though some have argued there is not much evidence to support these changes
taking place (McKinnon et al., 2008). This kind of reorganisation would trade off the economies of scale
from fewer production points against the high costs of transport (including unreliability).
Selecting a reliability metric
The discussion at the roundtable suggested there was no single correct metric for reliability to apply
in CBA. Any approach taken would either have to be highly disaggregated to reflect travellers having
different information and trip purposes, or must involve strong homogenising assumptions about
information, preferences and behaviour.
In practice, the metric currently typically applied in CBA is the variability (generally measured by
the standard deviation) of a distribution of origin-destination travel times that is implicitly known to the
traveller ex-ante. The traveller selects her departure time in advance (when the travel-time distribution is
estimated or “observed” by her). In the current practice, the mode, route/service and the
origin/destination of the trip are assumed fixed. No approaches to valuation or demand forecasting are
yet available that account for (some) travellers incorporating real-time “unexpected” reliability
information (while the trip is underway) into their route choices. The traveller may use information about

QUANTIFYING THE SOCIO-ECONOMIC BENEFITS OF TRANSPORT — © OECD/ITF 2017


22 – 1. IMPROVING TRANSPORT COST-BENFIT ANALYSIS: OVERVIEW AND FINDINGS

“expected” variations (time of year or day) to choose their departure time but once she is “on the road”,
is assumed to be an otherwise uninformed and passive recipient of travel-time outcomes.
Valuing reliability
Fosgerau (Chapter 2) outlines a theoretical structure for the rational decision making of a traveller
faced with an uncertain travel time for her upcoming trip. The models are in the scheduling model
tradition of Vickery (1969) and Small (1982) where utility from the trip depends on both the mean and
standard deviation of travel time. This definition is important as it shows how the traveller will value
reductions in this variation, even where the expected travel time (i.e. the mean travel time) is unchanged,
so the reliability measure is conceptually distinct from travel-time savings captured elsewhere in CBA.
The two variants are argued to be applicable to different types of trips:
•

The step model, where the traveller attributes a high utility to being at the destination by a
specific hour (but not otherwise), is relevant particularly for travel to employment with a fixed
start time or travel to appointments. Travellers described by this model will be motivated to
include a travel-time margin by departing earlier than would be implied by the average travel
time and their preferred arrival time.

•

The slope model, where travellers have more gradually shifting preferences to be at the
destination rather than the origin, is more relevant to travellers for whom the specific arrival
time is not critical, such as leisure travel or travel to employment with a flexible starting time.
Travellers described by this model would not incorporate a travel-time margin when deciding
their departure time.

Mathematically, each alternative model implies a different metric of variability that should be used
for valuation. The step model gives a valuation in terms of standard deviation (and Fosgerau (Chapter 2)
shows that the model is compatible with several alternative measures of dispersion, such as the mean

lateness, if the shape of the travel-time distribution is fixed). The slope model by contrast leads to a
valuation in terms of variance. This distinction can be important in the practical application, as variance
is simpler to sum across composite links in a journey to get an aggregate variability measure.
Valuations for average travel time and travel-time reliability are generated by fitting model equations
(either in structural form or reduced form) to a dataset. These are typically thought of as the costs
internalised by the traveller themselves, but other parties (e.g. a fellow meeting participant) may also
incur costs (Fosgerau et al., 2014). The two types of datasets that can be used in the empirical estimation
are stated preference (SP) drawn from survey responses to choice exercises and revealed preference (RP)
data drawn from observed travel outcomes. These approaches have strengths and weaknesses that were
debated at the roundtable (Table 1.1.).
Fosgerau (Chapter 2) highlights some fundamental challenges to the rationality assumptions
underlying the scheduling models when they are used to estimate the value of reliability with SP data.
While these were acknowledged, not all roundtable participants agreed with his suggestion that SP-based
valuation should be abandoned in favour of RP methods. What participants did agree on is that RP
methods should be pursued with renewed vigour as the amount of high quality RP data available to
researchers expands rapidly with the growth of smartcard public transport ticketing systems, GPS units in
vehicles and smartphone location data. There may also be opportunities to draw on the strengths of both
data sources through combined SP and RP estimation (Ben-Akiva and Morikawa, 1990).

QUANTIFYING THE SOCIO-ECONOMIC BENEFITS OF TRANSPORT — © OECD/ITF 2017


1. IMPROVING TRANSPORT COST-BENFIT ANALYSIS: OVERVIEW AND FINDINGS –

23

Table 1.1. Approaches to valuing reliability benefits
Approach

Strengths

Weaknesses

Stated
• Respondents may not be able to process the questions
• Relatively low cost to gain large sample
Preference
effectively to make choices reflecting true behaviour
• Control over the range of variation
(SP)
• Different valuations for reliability depending on
• Repeat observation of each respondent allows
whether structural or reduced-form estimation
control of personal/local influence on choices
• Very difficult to use for freight sector – which part of
• Can frame questions to align with assumptions
the supply chain to interview?
(e.g. ex-ante known distribution)
Revealed
• Observing real behaviours and decisions
Preference
• More comprehensive coverage
(RP)
• Data increasingly available

• Requires huge amounts of data (and processing power)
to generate valuation results
• Can be difficult to delineate the effects of average
travel time and variability on travel choices

• Feasible to provide valuations for freight sector
• Can be difficult to introduce monetary dimension
(though these may understate true values)
Source: ITF based on roundtable discussion and Fosgerau (Chapter 2).

Extending the valuation approach from the case of passenger car travel is likely to be achievable with
further research. In the case of public transport service reliability it may be that, as with average travel
time, travellers may experience disproportionate disutility from variation in the components of the
journey (wait time, in-vehicle time) and conditions (seating versus standing). For instance, there may be
greater disutility from unexpectedly standing on a bus for 10 minutes, compared with 10 minutes extra
on-board in a seat. Valuation for freight trips was seen by some participants as actually easier to model
and estimate, given the incentives are clear-cut. However, the issues with freight are perhaps more
fundamental. Freight operators are likely to be well informed about levels of variability, and so are likely
to have adapted their operations in advance of the “marginal trip” being considered (in either an SP or
RP context). The costs of unreliability (and hence the benefits from unreliability) will come more from
the reorganisation of operations, the changes in fleet and staff levels, etc. that will be undertaken to best
provide the required level of service to their customers. The marginal CBA framework will not easily be
extended to capture such costs, yet they are likely to be large.
Forecasting reliability
Coherently measuring, let alone forecasting, the ex-ante travel-time distributions faced by road users
between point A and point B is challenging. As described above, for any departure time on any given
day, each traveller would have their own ex-ante travel-time expectation, which might be a range or a
point estimate, is likely to vary considerably from person to person, and is not observable. Ex post, a
single travel time can be observed from number-plate recognition systems or loop counters.5 However, in
the absence of any empirical evidence to develop a theoretical model of the formation of expectations,
any approach needs to relate the real observed data to the expectations of the travellers. The approach of
Kouwenhoven and Warffemius (Chapter 3) is to assume travellers form a time expectation based on
times observed at the same time of day in the recent past (and even incorporating values from the near
future). The ex-ante unreliability perceived by the traveller is assumed to align with the ex-post traveltime distribution (around this expected value).
In developing ex-post distributions in 15-minute time slices Kouwenhoven and Warffemius (Chapter

3) exclude selected travel-time observations as “extreme event” outliers as recommended by de Jong and
Bliemer (2015) (Figure 1.4). This approach polarised participants: should extreme events be captured by
a separate valuation and forecasting framework or within a single framework? In the Dutch case, the
argument is that the valuation study, which used SP surveys, did not cover extreme delays. Moreover, the

QUANTIFYING THE SOCIO-ECONOMIC BENEFITS OF TRANSPORT — © OECD/ITF 2017